Interactive web map for IPL match data from 04/18/2008 to 05/20/2016. The interactive map can be found at http://rpubs.com/ysharc/184185
- Download IPL match data from http://cricsheet.org/downloads/ipl_csv.zip
- Extract it in any folder of your choice and run collect.py from the same folder (if running from a different location change the path of the files in collect.py)
- Now for plotting the data run plot.R in RStudio to view the map instantly (if running in the R console, execute the next command to export it to a html file) 1. library(htmlwidgets) 2. saveWidget(map, file="map.html")
##Quick Look
- Map starts with an overview of all IPL matches played until now.
- Select a subgroup to view their respective cluster. (For example kolkatta_wins in the following image)
- Hover mouse over a cluster point to see the area it covers
- zoom in to see the spread of the group
- At the final zoom level click the cluster to display a spiderweb of all the matches that took place in it.
- The starting point of the spiral contains early day matches, where as the last point contains the latest match
The interactive map creation can be summarized by the collection and plot parts. The overview of the processes and code are explained below
######Data Collection Download the Total IPL match data from http://cricsheet.org/downloads/ipl_csv.zip . The archive contains a csv file for every match of the event. Each file typically contains the following structure
| version | 1 | '' |
|---|---|---|
| info | team | Delhi Daredevils |
| info | team | Sunrisers Hyderabad |
| info | gender | male |
| info | season | 2016 |
| info | date | 5/20/2016 |
| info | competition | Indian Premier League |
| info | match_number | 52 |
| info | venue | Shaheed Veer Narayan Singh International Stadium |
| info | city | Raipur |
| info | toss_winner | Delhi Daredevils |
| info | toss_decision | field |
| info | player_of_match | KK Nair |
| info | umpire | A Nand Kishore |
| info | umpire | BNJ Oxenford |
| info | reserve_umpire | YC Barde |
| info | tv_umpire | VK Sharma |
| info | match_referee | S Chaturvedi |
| info | winner | Delhi Daredevils |
| info | winner_wickets | 6 |
The remaining lines contain ball to ball info of the two innings, but the columns are inconsistent. For example a normal ball contains 8 columns where as a wicket contains 11 columns. Using glob library extract all the names of the csv files to access
file_names = glob.glob('*.csv')
Next using the csv library open an input file, create a writer object and create the header
with open('input_data.csv','wb') as input_file:
writer = csv.writer(input_file,quoting=csv.QUOTE_ALL)
writer.writerow(['Match','Season','date','venue','city','lat','lon','Winner','sf'])
With the input_file open iterate over every file name in the list file_names and extract info using the pandas library. For example extract the date of each match as shown below
for filename in file_names:
df = pd.read_csv(filename, warn_bad_lines=False, error_bad_lines=False)
date = df.loc[df['version']=='date'].iloc[0,1]
Extract whatever features you want to plot on the map, here in this example the match,date and winner are plotted. The next important step is to collect co-ordinates of the location using any geocoding API or webscraping. Here in this example the nominatim search is used to query for co-ordinates(more info on parameters at http://wiki.openstreetmap.org/wiki/Nominatim#Parameters).
Using the requests and beautifulsoup libraries collect and scrape the data you need.
response = session.get('http://nominatim.openstreetmap.org/search.php?q='+venue.replace(' ','+')+'&polygon=1&viewbox=')
soup = bs4.BeautifulSoup(response.text,'html.parser')
required_element = str(soup.select('script'))
phrase = re.compile('nominatim_results(.*?)\[(.*?)\]')
ans = phrase.search(required_element)
lat_lon_str = ans.group()
lat_pattern = re.compile(ur'lat": "(.*?)"')
lat = lat_pattern.search(lat_lon_str)
lon_pattern = re.compile(ur'lon": "(.*?)"')
lon = lon_pattern.search(lat_lon_str)
Now that you have collected all the data needed, update our input file with the info gathered
writer.writerow([match,season,date,venue,city,venue_lat_lon[(venue,city)][0],venue_lat_lon[(venue,city)][1],winner,sf])
Finally you have the input_data.csv file created with all the info you wanted, now moving on to the plot process.
In this process we create interactive web maps using leaflet and htmltools packages of R. The first step is to read in the data we created in the previous process
IPL = read.csv("input_data.csv")
Create subsets of the dataframe for analyzing the win locations of each team. For example
chennai_wins <- IPL[which(IPL$Winner =='Chennai Super Kings'),]
Next create icons for each team to display on the map
teamIcons <- iconList(
csk = makeIcon("csk.png", iconWidth=36,iconHeight=36),
dc = makeIcon("dc.png", iconWidth=36,iconHeight=36),
dd = makeIcon("dd.png", iconWidth=36,iconHeight=36),
gl = makeIcon("gl.png", iconWidth=36,iconHeight=36),
kxp = makeIcon("kxp.png", iconWidth=36,iconHeight=36),
ktk = makeIcon("ktk.png", iconWidth=36,iconHeight=36),
kkr = makeIcon("kkr.png", iconWidth=36,iconHeight=36),
mi = makeIcon("mi.png", iconWidth=36,iconHeight=36),
pw = makeIcon("pw.png", iconWidth=36,iconHeight=36),
rr = makeIcon("rr.png", iconWidth=36,iconHeight=36),
rps = makeIcon("rps.png", iconWidth=36,iconHeight=36),
rcb = makeIcon("rcb.png", iconWidth=36,iconHeight=36),
sh = makeIcon("srh.png", iconWidth=36,iconHeight=36),
t = makeIcon("draw.png", iconWidth=36,iconHeight=36)
)
Now initialize the map and set the basemaps.
map <- leaflet(IPL) %>%
#overlay groups
addTiles(group = "OSM (default)") %>%
addProviderTiles("Esri.WorldGrayCanvas", group = "Esri.WorldGrayCanvas") %>%
Next add markers and cluster options to the data and finally add overlay Controls.
addMarkers(data=IPL,icon=~teamIcons[IPL$sf], group = "IPL",popup = ~htmlEscape(paste(Match,"***",date,"***","Winner: ",Winner)),clusterOptions = markerClusterOptions()) %>%
addMarkers(data=Matches_that_tied,icon=~teamIcons[Matches_that_tied$sf], group = "Matches_that_tied",popup = ~htmlEscape(paste(Match,"***",date,"***","Winner: ",Winner)),clusterOptions = markerClusterOptions()) %>%
# Layers control
addLayersControl(
overlayGroups = c("OSM (default)", "Esri.WorldGrayCanvas"),
baseGroups = c("IPL","Matches_that_tied"),
options = layersControlOptions(collapsed = FALSE)
)
#References